Class D amps seem poised to take over. Then what?


I am certainly biased by my lifetime final amp being a Class D. But I know that after 30+ years of development, Class D seems to be on a high plain. I know there are now many, many companies focusing on Class D and, maybe, a good handful already as good as it gets. My Class D amp is as smooth and beautifully musical as a great tube amp and as punchy and detailed as a great SS amp. I am satisfied and done with my search. A class D amp has effectively taken me off the amp merry-go-round. It’s about time after 50 years. And, for me, this Class D is a milestone. Will all other classes of amps fade away?
mglik
...yes, but is Class D stable into 2 ohms?
Sadly not, like good Class-A/B's can be using BJT (bi-polar) output stages, especially in current delivery.

Cheers George


jbhiller
I’ve tried a couple class D units. They were good but not great to me. I’m game to try some more.


If you can get hold of a Technics SE-R1 flagship or at least it’s cheaper integrated the SU-R1000
The SE-R1 addresses the "Achilles Heel" of Class-D, (ultra sonic switching noise down into the audio band) Technics takes it up 3 x higher and so filters it up 3x higher also, which results in 3 x less switching noise in the audio band.
It’s been said the SU-R1000 integrated does it also, but not confirmed by Techinics.

These are the only Class-D’s so far to take full advantage of the GaN Technology’s far higher ability to be able switch 3 x higher.
Other manufactures using GaN are only half exploiting it, and still using the old 3 x lower switching frequency, like all other Class-D’s do since the 90’s, so really not much advantage using GaN if not used like Technics have.

This top pic is what the switching noise looks like on the output of a Class-D going to the speakers a 10khz square. https://ibb.co/qBnNwVJ
Now Stereophile use an "AP (Audio Precisions) temporary test bench filter" on the output to give the illusion of a cleaner bottom pic, and they never give the top pic anymore, this is done in my view as to to not freak potential; buyers out.

Cheers George
Grannyring,
I am 100% sure it will sound different. Every single thing you do changes the sound including jacks, wire, solder, chassis vibrations, etc. But, yes, they are using the exact same modules (except that Rouge is using mono modules and Legacy is using one half of a stereo module (exactly the same thing). Yes, you can see pics of the inside of the Legacy amp that verify this fact. Legacy are not doing anything to the modules....they do not touch them......it is quite clear from the pics. They do have a separate ground wire running from each module and they do scrap away the anodizing where they ground (standard practices). Basically, they are stock IceEdge modules in a box.....same as Rouge Audio. What I am 100 % sure of is that with my mods either amp will sound way, way better.....but still slightly different from each other.

https://systems.audiogon.com/systems/5078
Pic #11 shows the best interior shot. Click on the pic again to make it larger.....you can clearly see the amp channel on the top of each module is not hooked up in any way.  Totally stock stereo modules using just one channel for more power and better separation (same as using mono modules).

George,
Wrong, many class D amps are fine with 2 ohms (stable and otherwise) and one of my customers is driving 2.4 ohm Thiel 3.7s ......Fremer drove his 1.4 ohm Wilson Alexx with the IceEdge module without a hitch. Time to tell the truth, dude! Yes, some class D amps are not designed with big enough power supplies to drive 2 ohms or less. After all, how many speakers go down to 2 ohms or less.....mostly the big expensive ones that only the 1% can afford and with 97+db speakers like the incredible sounding Lii Audio drivers.....those big expensive boxes are now boat anchors.
yes, but is Class D stable into 2 ohms?


Yes, they are stable into 1ohm, the H20 amp that Henry used to make was used with the apogee scintilla.  

FACT!!!!
No way can ANY Class-D compete with doubling of wattage into impedances 8 > 4 > 2 and 1ohm (current ability) like BJT amps Gryphon Antillion, bigger Krell, D’Agostino Boulder, etc etc etc can

Class-D "may" not blow up, because of severe current limiting at 2 and 1ohm but they can’t drive into those low impedances like those BJT above amps can.
Anyone who says they can is just shilling for sales.

A friend Edgar Kramer, editor/owner of Soundsatge Australia has the Wilson Alexia MkII, and no Class-D he’s tried and reviewed and I’ve heard at his house has sounded any where near as good on them, as his Gryphon Antilion does, especially in the bass where those Alexia’s are 0.9ohm loading.
I will also say the Technics SE-R1 which I rave about is also in that same situation, as GaN Fet like Mosfets output stages can’t do current like BJT’s can into low impedances, 1 or 2ohms

Cheers George
A lot of scintilla owners went from the big krell amps to the H2O and said it was way better.
The Merrill amps double power to 2 ohms.....who cares? about one ohm except for those that have overpriced big speakers that only the 1% can afford?  You really need to try the Lii drivers on an open baffle.  You can use 10 watts and play way over 100db.  Those Wilson's etc. are boat anchors.  You buy a $100K speaker and you have to buy a $40K amplifier to run them....crazy.  A pair of 99db Lii 10 inch full range drivers are $1300 a pair delivered.  Make an open baffle with wings and you have flat to 40hz bass.....and then in a few months (after you have picked your jaw off the floor) you add a supertweeter....then later add some woofs....and you have better sound than a Wilson big heavy boat anchor.
My AGD Audion’s drive Wilson’s effortlessly even the large XLF’s. I own Wilson’s no issue at all. And the $7500 AGD's sound better than some 40-50k amps.


Guarantee put a Gryphon Antilion stereo or monoblocks on them and you'll change your mind in an instant.
And if you don't there's something seriously wrong.

Cheers George
I said the AGD's sound better than SOME 40-50k amps. It's also not my opinion alone i have had many bring their mega buck amps into my system and most are shocked at the SQ of the AGD's most leave scratching their heads in disbelief.

By your comment you must have heard the AGD's so what's YOUR take on the SQ?

Don’t have to, the AGD is just GaN based Class-D slid into a "bogus" empty tube envelope, which uses (horror) the base pins for all it’s contacts 🤦‍♂️

And there no way the AGD can do what say the Gryphon Antilion can do into the Alexia, Alex, or XLF’s bass, if you do own any of these, they go down to 0.9ohm EPDR (Equivalent peak dissipation resistance). It’s a measurable fact.

Don’t get me wrong, I never said it won’t sound any good, just not the very best for those speakers. And if I had the money for them I’d have the money for the best amp also.

Cheers George
Keep your Dingos away from him
What does that even mean?
@georgehifi - I read the book written by the Chamberlain's lawyer, and learnt a lot about Dingos, they are most closely related (genetically) to the timber wolf. Highly intelligent is accurate. I've only ever seen them in zoos, never in the wild, even when I drove around Oz.
...yes, but is Class D stable into 2 ohms?
Yes, depending what most meanings of the word 'stable' are. Yes, a class D will not go into oscillation when presented with a 2 ohm load (instability can result in oscillation- so that's one meaning of the word...). Yes, a class D amp can easily double power into 2 ohms or even 1 ohm, since its actual output impedance is probably only a few milliohms (our Beta production amps have an output impedance of about 10 milliohms).


With an output impedance like that, of course an amp can double power. But we also have to be clear about what limitations exist. The limitations are the ability of the heatsinks to get rid of the additional heat, the current available in the power supply and the ability of the output devices to handle that much current. So most class D amps might be able to double power into 2 ohms, but maybe not at full power.


One limitation with GaNFETs is the heatsink- because their appearance has a lot in common with a postage stamp, there's not a lot of methods available to couple them efficiently to a heatsink. And that heatsink has to be very effective at moving heat away from the device. This is (IME) the biggest limitation of GaNFETs; on the case of our design, the heatsink simply isn't fast enough to move all the heat away from the devices if operating at 2 ohms to full power (which would be 800 watts). The output devices are rated for the current, no worries there and its no problem putting in a power transformer that has enough current.


But are they stable? Certainly! For that matter so are any of our tube OTL amplifiers (in that they too will not oscillate when presented with an adverse load) although they certainly won't double power. So what this question really points to is the ability of the amplifier to behave as a voltage source (able to put out the same voltage regardless of load) and so long as you don't ask it to make full power into the load, almost any class D will do that like a walk in the park.


But at the same time its important to know that  **ANY** amplifier driving a low impedance like that will not really be able to strut its stuff. All amplifiers make higher distortion when driving lower impedances; from a high end audio point of view where its all about getting closer to the music, this is the Thing You Don't Do if you want your stereo to sound like real music- the additional distortion will manifest as harshness and brightness with a loss of detail.
George
What you posted shows lack of knowledge. Your talking about products that you have zero knowledge about, why say anything at all?

I hate it when people talk about subjects they know nothing about.
Never entirely but for the most part, class D is too attractive an option to audio manufacturers with it's cool running efficiency, compact form factor, etc.
Properly implemented, it's increasingly recognized as being as good or better than the alternatives, despite myths to the contrary.

I have been selling audio for along time now and we have heard this type of rhetoric before. A few decades back a person buys a cd player and claims to the world that the turntable will now be obsolete, more recently another person bought a streaming device and declared to the world once again the turntable will be obsolete. The names change but the story its still the same. Very seldom do our customers every make a request for a certain class of amplifier, instead  their decisions are based on how it sounds with their equipment and room. Class D amps have made themselves known in the huge audio equipment landscape, are they going to make everything else obsolete? History says no. For us the problems we have encountered with class D is reliability and repair
I hate it when people talk about subjects they know nothing about.
You need to hate yourself then, as you have no idea about the performance of Class-D even when using GaN Fets into 8 4 2 and 1ohm at it’s "current ability" (wattage doubling) at these loadings. And how those Wilsons demand it to get the best out of them.
Next you’ll say an OTL of Ralphs can drive those Wilsons to their best properly, because they too will stay stable into them and not oscillate.🤷‍♂️
@georgehifi - wouldn’t it depend on the original rating at 8ohms, the power supply rating that feeds the class D modules?

I’m not pretending to know, I’m just posing the question: if the power supply is engineered to supply enough current at 1ohms impedance seen by the amplifier, where is the limitation?
Could it not be possible, to double to 1ohm?
Is it a lack of efficientcy, I mean there’s less loss by creating heat.

It might be a lengthy topic to answer, if you know of an article I could read about it perhaps?
WOW! You know what my system sounds like yet you have never heard it.

Now your a mind reader knowing what i will say next. Bravo!

I was in the industry for many decades, building and designing, amps, pre amps and speakers so it's easy for me to tell who's winging it and those who know their stuff, keep winging it george i know better.


@georgehifi - to be clear here, I saw the post after mine (above this) and I wasn't be combative, I was giving you the benefit of the doubt.

I know a bit more than Ohms laws, so I have genuine interest as to the limitations. Which there very well may be.
@georgehifi - wouldn’t it depend on the original rating at 8ohms, the power supply rating that feeds the class D modules?

No, although it to needs to be beefy, look at any Class-D that has been independently wattage tested into 8ohm 4ohm 2ohm regardless of powersupplies (most won’t get test into 2ohms), they are fets, and fets can’t do doubling of 8 to 4 to 2ohms (current) like complimentary bi-polars (bjt) can do like in the the Gryphon Antillion, bigger Krell, D’Agostino Boulder, etc etc etc

Also look at my post in the  "Pass X260.8 vs Bryston 7B3" thread

Cheers George
rh67
Would you agree that the actual GANfet used in AGD amps is proprietary? The comparison has been made to Nelson Pass acquiring all available SITs. Although Nelson does not hold the patent on SITs as does Alberto on the AGD’s GANfet.
One reason I bought AGD is because I was convinced that they stood apart from all other Class D.
That all the others used variations of existing technology (chips).
earlflynn, thanks for the response. I totally understand passion,and brand loyalty.  I'm a motorcycle enthusiast. I ride Victory bikes. I was also in an MC where everyone else in the club rode HD. While we used to bust each other's stones, it never got to the point of hostility.  I guess when you can only see written words,sometimes what is meant to be passion,comes off as pompous.  
You are correct.

Also to date they do stand apart design wise Merrill does also atmosphere from what i'm reading has spent many hours designing their Class D offering which i'm sure will be a great sounding amp.

But in the end it's the SQ, one can have the best parts but if you do not know how to maximize what you have best parts means little.

You know as i do from experience and living with the AGD's that they are great sounding, Alberto has taken advantage of what he has at his disposal. It does not take long after talking to him that he is the real deal and a very intelligent person when it comes to design and parts implementation. His upper end amps have a clear cover so one can see his design also in the so called tubes. He is very proud and deservedly so of his creations.

Some talk negative about the amps but do they have any experience with them? It's like me telling you how your care drives even though i have never been in the make or model, it also takes time to familiar yourself with a product to maximize it's potential.

It amazes me how some can form an opinion out of the air.


It amazes me how someone can say a Class-D like the AGD has no limitations into one of the hardest know speakers to drive, and for it to sound it's best.


They can come close if good, like the BJT ones I mentioned.,
Where many Class-D go backwards! instead of trying to double, if independently measured into 2ohm from the 4ohm wattage, that shows just how current limited they are.
Everyone has an opinion - which is good - not everyone agrees with another persons opinion -  you all know my opinion on this subject matter.

Happy Listening. 
Everyone has an opinion - which is good - not everyone agrees with another persons opinion


Difference is, these are measured facts, and not my opinion.
And if one dismisses them then this is what they believe, 
https://i.dailymail.co.uk/i/pix/scaled/2013/12/18/article-0-1A2B262D00000578-818_636x382.jpg
as every decent piece of audio gear is designed by the laws of electronic design 
Next you’ll say an OTL of Ralphs can drive those Wilsons to their best properly, because they too will stay stable into them and not oscillate.🤷‍♂️
The sales manager of Wilson had our amps for many years. FWIW our OTLs are inherently stable and will not oscillate with any load or input signal condition. 
they are fets, and fets can’t do doubling of 8 to 4 to 2ohms (current) like complimentary bi-polars (bjt) can do like in the the Gryphon Antillion, bigger Krell, D’Agostino Boulder, etc etc etc
This one gave me a good chuckle- its so ridiculously false as to be funny! This statement ignores a simple fact about GaNFETs which is their ON resistance is one of the lowest values of any semiconductor.
One reason I bought AGD is because I was convinced that they stood apart from all other Class D.
That all the others used variations of existing technology (chips).
It certainly helps when designing a switching amplifier to have also participated in related patents to switching transistors as Alberto has! But if you want to know why his amplifier sounds right, take a look at the distortion signature he shows on his website. If you look at the values you'll see that its not particularly low, but what is important (and IME arguably more important that *how much* distortion you have) is that the signature is what you want to see to prevent the amp sounding harsh. Here's the link
https://agdproduction.com/audion/
scroll down and click on either of the THD images. You'll see that the 2nd and 3rd harmonics predominate. These harmonics are nearly inaudible to the human ear and because they are so prominent, they mask the higher orders (which otherwise are audible usually as harshness and brightness). This distortion signature looks very similar to that of our OTLs (although we show more 3rd and less 2nd due to our differential design). This type of distortion signature will cause **any** amplifier to sound smooth and organic. The lower the distortion the more detailed the amp will be without also being bright.


I've said many times that the differences we hear in amplifiers has more to do with what distortion we can hear and what distortion that we can't. This is more important than output impedance because the ear has a tipping point where tonality generated by distortion gets more attention than actual frequency response.


Class D amps can have a signature like this because the errors (non-linearities) in the encoding scheme and the dead time requirements (if handled correctly) tend to result in lower ordered harmonic production. (There is more to it than that of course...) In a nutshell this will cause them to sound like a state of the art tube amplifier that acts as a virtually perfect voltage source.


If you are pragmatic as a designer you know that you can't make an amp with no distortion :)  The savvy designer will thus see to it that the distortion that is present is innocuous and this is a good example of how that is done.

georgehifi
Next you’ll say an OTL of Ralphs can drive those Wilsons to their best properly, because they too will stay stable into them and not oscillate.🤷‍♂️
atmasphere
The sales manager of Wilson had our amps for many years. FWIW our OTLs are inherently stable and will not oscillate with any load or input signal condition.

Dream on sunshine, were talking about driving the Wilsons mentioned here.
Trust you to try to skew it around to make out that your OTL amps could even come close to driving these, the hardest of Wilsons to drive
(Always looking for the self promoting angle)
Keep your Dingos away from him
What does that even mean?


Anger issues, I'm guessing.
@invalid.FYI.  These guys claim that wattage doubles (close enough) at 4ohms. 

Pathos Classic MK III
Specifications

Type: Integrated amplifier in Class A/AB, hybrid circuit with input tube stage, bridgeable
Output Power: 2 x 70 W @ 8 Ohm, 2 x 130 W @ 4 Ohm, 180 W @ 8 Ohm in bridged mode (mono)
Inputs: 1 balanced XLR, 4 unbalanced RCA
Frequency response: 2 Hz - 100 kHz
THD: <0.05%
Signal/noise ratio: 90 dB
Input impedance: 100 kOhm


Pathos LOGOS

The Logos uses MOSFETs in the output stage, known for their tube-like transfer characteristics, in a design that attempts to come sonically as close as possible to Pathos’ patented InPol circuitry. InPol is the only amplifier design, hybrid or solid-state, that gives the full glory of a tube amplifier but with control. Look at the numbers: 110 watts per channel into 8 ohms, and an amazing 220 watts into 4 ohms, with full bandwidth


Output Power: 2 x 70 W @ 8 Ohm, 2 x 130 W @ 4 Ohm


It wasn’t mentioned
Let see what happens into 2ohm un-bridged. (bet they can’t handle it) and those Wilson’s mentioned hit 0.9ohm!!!.

Cheers George
You keep mentioning Wilson's as being hard to drive, where does this come from? I have owned Wilson's from their beginning starting with the original Wamm's and i have never found them difficult to drive even using 300B tube amps. It seems that you lack personal experience with Wilson. https://www.stereophile.com/content/wilson-audio-specialties-alexandria-xlf-loudspeaker-measurements
Things always get better....I remember when the new craze was introduced ...CD.  The sound was awful, until it got better.   If you live long enough there will also be something better waiting in the wings.  Digital can be excellent, but excellence comes at a price as does everything.
You keep mentioning Wilson’s as being hard to drive, where does this come from?

For someone who "says" they’ve owned Wilsons, you are not very au fait or even interested in how they measure or behave, in particular the loading they present for amps to see. And they have many models that are in that same low EPDR loading bracket.
From HiFi News Lab Report
Impedance phase is quite well controlled but still the EPDR falls to a low of 1.2ohm/67Hz, with further minima at 129Hz (1.8ohm) and 410Hz (1.9ohm). So the Alexia 2 presents a challenging load to the amplifier, albeit not so much as its predecessor Mk1 (min EPDR 0.9ohm!!!!).

These companies that claim double wattage when impedence is halved don't give the actual 8ohm measurement, which is usually a lot higher wattage than their spec sheet has listed.
These companies that claim double wattage when impedence is halved don’t give the actual 8ohm measurement

Usually if they don’t, they are hiding something. Stereophile does for Class-D if you read the Lab Tests on them and that’s usually is in the negative.
But even Stereophile doesn’t usually do 2ohm testing on Class-D’s as they are usually pitiful, when compared to the good linear solid BJT state amps for putting out big (doubling watts) current.

Cheers George
Even the big high bias amps built with BJT's can't do it. When stereophile measures them their 8ohm full power measurements are a lot higher than the spec.
I don't think there is an amp made that can direct drive the apogee full range, the midrange when direct driven is .14 ohms
Even the big high bias amps built with BJT's can't do it.

Oh dear🤦‍♂️, to say that an amp doubles, is just a "saying", maybe it should read "close to doubling", but hey, I didn't make up the saying.

No! amp can  perfectly double it wattage from 8  to 4 to 2 ohms, that would be a furphy, and would be like saying perpetual motion is a reality.
I can second the comments by those who have/have heard AGD amplifiers.
My bias towards Class D has been negative -my Bladelius Ask has a ‘green’ mode using a Class D module, which is flat and harsh compared to the ‘performance mode’ which is ClassA/AB and very good.
New speakers in another system are higher efficiency ( Viking Acoustics Berlin), and the LSA Statement (hybrid integrated) driving them sounds wonderful, yet is simply too noisy. I hear the tube hiss at lower volume and can’t ‘unhear’ it. My search for a replacement led me to the AGD Audion, which I had the pleasure to listen to at Ocean Way Audio in LA. I found everything stated about these amps, also in this thread and particularly by @atmasphere, confirmed. These are superb amps. Tube like sound and dead quiet. It does of course help when one of the most accomplished recording engineers in the industry endorses a product. In my mind, this is not a mere variation of a theme, it is a discrete innovation when you consider form factor and performance.
I ordered a pair…